Linear and nonlinear earthquake location approaches in a case study overview

Abstract

Several seismic events are calculated by an iterative, linear algorithm relocated in a nonlinear, grid-search with a maximum-likelihood hypocenter method. The same simultaneous events separately registered by regional and local seismological networks have also been used. A verified 1D (one dimensional) velocity crustal structure model is used in both methods and in all datasets for better comparisons. Some rock quarry blasts registered by the local receiver stations as ground truths are examined by both methods. This study's main objectives were: (i) to carry out a comparative examination of the error estimates in both techniques and (ii) to generalize location assessment of the permanent low-density regional network in Central Alborz. The regional network operates under Iranian Seismological Center and its seismic data are massively applied. To reach the research goals, the dataset–method correlation matrix and the location variables were statistically processed. It was shown that the estimated earthquake epicenters by the linear and nonlinear methods are approximately similar for well-located events in each network. For dense-proper distribution of the seismic stations, the elliptical error estimates have lower amounts and consequently the seismicity pattern is more reliable. The difference between subsurface seismicity patterns resulting from the above-mentioned approaches in the studied area is presented. Some depth calculation errors for the same recorded events are higher in the regional network. The correlation comparisons demonstrated that the nonlinear approach and the locally-constrained data may be closer to reality and more convincing. The nonlinear application is more plausible due to all of the possible solutions and less elliptical error estimates under the Gaussian distribution assumption. Subsequently, the combination of the nonlinear method, proper location criteria, and appropriate crustal velocity structure can effectively improve some of the earthquake location estimates in the regional catalogue. As a result, this procedure can be useful in seismicity concept evaluations.